GitLab

hpa-tftpd will not start.

table.

count.

and its good not to use the same variable name for both!

patch to avoid vlan corruption.

for now.

switch interface types.

what it appears to be - it's a num, not an int.

since that is now done automatically. I do mention it, but in setup
where other little misc bits are mentioned (stuff that boss-install
does).

case there are any dependancies.

installation instructions tell them not to install the ports
collection or any packages. Include instructions on how to get our
approved version of the ports tree.

Remove the text about updating to -STABLE, as that is no longer our
recommended procedure.

blend triangles can be plotted overlaid on the all-cameras composite figures.
Add camera_data Makefile targets to do the composite figures.

software.

	source tb_compat.tcl
	set ns [new Simulator]

	tb-elab-in-elab 1

	namespace eval TBCOMPAT {
	    set elabinelab_maxpcs 1
	    set elabinelab_nodeos(myboss) FBSD47-STD
	    set elabinelab_nodeos(myops) FBSD47-STD
	    set elabinelab_nodeos(mypc1) FBSD47-STD
	}
	$ns run

work inside ElabInElab.

The crux of this is to emulate what switchmac does when incorporating nodes
into the testbed. Rather then using snmpit to ask the switches for MAC info,
we can just ask outer emulab via the proxy to get that same info from the
DB. Note, there are no changes to the newnode MFS; it will boot and happily
report it after launching teachswitch, but those packets are harmless to the situation.

Another wrinkle is that we need the outer emulab to tell us which
interfaces are control and which are experimental, since inside there
is no real way to determine that. I did this by overloading the final
field in the return from switchmac, and using that to override the
"role" setting that utils/newnode would normally choose on its own.

There is another new routine in the XMLRPC server that is intended to
be used for bypassing the newnode path. This is not hooked into
anything yet, but the intent is that rather then using the web
interface to "add" nodes, we just return everything needed to seed the
new_nodes and new_interfaces table, and then run newnodes directly.

I've sucesfully used this core to 'evolve' a sorted list of 18
elements. It takes negligible time, and ~100 generations.

Right now, I am only using mutation, not crossover. To do crossover,
I'm going to have to start studying permutation encoding.

mysqld hangs that cause the entire system to grind to a halt. The
basic theory of operation is like this:

* Once a minute fork a child (protected by a 60 second timeout) to
  connect to the DB and issue a simple query. If the child can access
  the DB okay, it exits with a zero status.

* If the alarm fires, the child is killed. This indicates that mysqld
  is no longer responding in a reasonable amount of time (60 seconds).
  We shift into trying to restart mysqld:

     * Send mysqld a TERM. Wait for 30 seconds.

     * Try query again; typically, the situation will not have changed one
       bit, but I do it anyway.

     * If mysqld was running, send it a kill -9. Wait for 15 seconds.

     * Start mysqld. Wait for 5 seconds.

     * Try query again. If query succeeds, we are done, and no one
       will have to deal with it Sunday morning at 6am (thanks Tim).

     * If query still fails, send email and give up trying to do fix
       anything. The daemon continues to query the DB once a minute;
       once the query succeeds (cause a human fixed things up), the
       daemon goes back into its normal mode (attempt to fix things
       next time it fails).

So, the problem is what happens when someone kills off mysqld for some
other reason. It may be that this daemon should only try to restart
mysqld if and only if, it actually killed a running mysqld. Comments?

kill fails.

that to keep files private to each subgroup.

mortals and robots can find the silly thing (otherwise it uses a very
silly name).

daemonize after making the connection to capture so an error is
reported if it cannot connect.

case.  Otherwise we don't pick up on experiment modifications.

some error handling, and put the output file sizes/mtimes in the status
file to make it easier to do a sanity check.

causing the select() to fail.

I still have some optimizations to perform.